Search

In the world fifty million peoples were affected by epileptic seizure disease. The brain activity of people is monitored by using Electroencephalography (EEG) Signals, which is utilized to identify the epileptic seizure diseases. In clinical field human master is important to measure and investigate the EEG to recognize normal signal and epileptic seizure signal. This technique is burns-through time, expensive, and the signal is additionally high in unpredictability. It is very crucial for accurately detect the seizure for antiepileptic drug treatment to prevent the future risk in the patients. In this paper Grey Wolf Optimization (GWO) based deep Recurrent Neural Network (RNN) approach is employed for accurate discovery of epileptic seizure in brain waves. The proposed seizure detection method has categorized into three steps. In the primary level the brain waves are decomposed into subbands by manipulating discrete wavelet transform. In the subsequent step, features namely line length, autocovariance and autocorrelation and periodogram, are extracted from the signal. In the following step GWO based deep recurrent neural network is applied to categorize seizure brain waves and normal brain waves from the selected features. The analysis of this method is performed using real time database and yields precision of 93.4%.The empirical scores exhibit that the existence of epileptic seizures in brain signals is effectively detected by this method.

In forensic odontology, the challenging task is to identity the decomposed and severely burnt corpse of individual person. In such a situation, dental records have been used as a prime tool for forensic identification. The main goal of this work, by comparing the analysis of contour shape extraction and texture feature extraction for both radiographic and photographic images, is used to identify a person. In this work, contourlet transform is used as a contour shape extraction; Local Binary Pattern (LBP), Center-Symmetric Local Binary Pattern (CS-LBP) are used as texture features. Both AM and PM images are used to identify the person more accurately by comparing different matching algorithms. In order to salvage better matching performance, Cumulative Matching Curve (CMC) is used for both radiographic and photographic images. Better matching is observed for radiographic images than photographic images by Hit rate performance metrics.

Electroencephalography (EEG) Signals are widely used to determine the brain disorders. The Electrical activity of human brain is recorded in the form of EEG signal. The abnormal Electrical activity of the human brain is called as epileptic seizure. In epilepsy patients, the seizure occurs at unpredictable times and it causes sudden death. Detection and Prediction of Epileptic seizure is performed by analyzing the EEG signal. The EEG signal of human brain is random in nature, hence detection of seizure in EEG signal is challenging task. Hardware implementation of Epileptic seizure detection system is necessary for real time applications. In this work an accurate approach is used to identify the Epileptic seizure and that is implemented in FPGA (Field Programmable Gate Array). The hardware implementation of epileptic seizure detection algorithm is done using Xilinx System generator tool. In the first step the EEG signal is extracted from the human brain and it is filtered by using Finite Impulse response (FIR) band pass filter. The band pass filter separates the EEG signal into delta, theta, alpha, beta and gamma brain rhythms. The band separated brain signal is modeled by linear prediction theory. In the next step features are extracted from the modeled EEG signal and the classification of normal or seizure signal is done by using Extreme Learning Machine (ELM) classifier. The EEG signals used in this paper were obtained from Epilepsy Center at the University of Bonn, Germany. The hardware architecture, Look up tables, resource utilization, Accuracy and power consumption of the algorithm is analyzed using xilinx zynq- 7000 all programmable soc.

Electroencephalogram (EEG) signal is widely adapted for monitoring epilepsy to rejuvenate the close-loop brain. Various conventional techniques are devised for identifying seizures that depend on visually analyzing EEG signals that are an expensive and complicated process if there is a rise in numbers of the channel. A new technique, namely Adaptive rag-Rider optimization algorithm (Adaptive rag-ROA) is presented to train deep-stacked autoencoder (Deep SAE) for discovering epileptic seizures. Initially, the EEG signals are given to the pre-processing module, in which noise has been removed by the bandpass filter. Then, the noise removed signals are provided as an input wherein the EEG is divided into various channels and each channel performs the extraction of features. Here, the features such as Taylor-based delta AMS, Holoentropy, fluctuation index, relative energy, tonal power ratio, spectral features, and linear prediction coefficient (LPC) are acquired from each channel. Furthermore, the Probabilistic principal component analysis (PPCA) is adapted to diminish the dimensionality of features. The obtained feature vectors are fed to Deep SAE for epileptic seizure recognition. The Deep SAE training is carried out with Adaptive rag-ROA that is devised by incorporating the Adaptive concept in rag-ROA. Thus, the output generated from the proposed Adaptive rag-ROA-based Deep SAE is adapted to detect seizures from EEG. The proposed Adaptive rag-ROA-based deep SAE outperformed revealing the highest accuracy of 91.5%, the sensitivity of 85.2%, and specificity of 86%.

Forensic Odontology is the evaluation of dental information that includes ante-mortem (AM) and post-mortem (PM) radiographs for the purpose of identifying person in some grave situations such as mass fatalities, natural disasters and terrorist attacks etc. One of the key issues in using dental images is that, although both the AM and the PM radiographs belong to the same person, there may be a mismatch between those radiographs due to the missing tooth in either of the radiographs. In such a case, the missing tooth in the radiograph has to be identified prior to the matching in order to achieve accurate identification of an individual. Thus an automatic algorithm for person identification in dental radiographs and photographs is a more challenging one at present. In this paper, texture based shape extraction algorithm is taken for analysis. Distance measures and classifier based approaches are the shape matching algorithm which is used to match both AM and PM images in order to obtain exact person identification. A novel approach has to be introduced for the extraction of the missing tooth, and subsequently each tooth in the radiograph is classified using k-NN classifier with Hu’s moment invariants as feature. Then each individual tooth is separated with pulp, enamel and dentine is applied to GLCM texture features. In this paper, a novel framework has been proposed to improve the identification performance. Moreover, the proposed algorithm achieves an overall accuracy of 98% than the existing approaches.

Forensic Odontology is a specific means for identifying a person in which deceased, and particularly in fatality incidents. The algorithm can be proposed to identify a person by comparing both postmortem (PM) and antemortem (AM) dental radiographs and photographs. This work aims to introduce a new mathematical algorithm for photographs in addition with radiographs. Isoperimetric graph partitioning method is used to extract the shape of dental images in forensic identification. Shape matching is done by comparing AM and PM dental images using both similarity and distance measures. Experimental results prove that the higher matching distance is observed by distance metric rather than similarity measures. The results of this algorithm show that a high hit rate is observed for distance based performance measures and it is well suited for forensic odontologist to identify a person.

In Recent years increasing attention is paid to welding defects monitoring in industries. An automatic system to extract and classify the welding defects in radiographic images is a main challenge in industries, because of the miniature welded region. In this paper an automatic method is proposed to extract the various welding defects in radiographic weld images. In the first stage the input radiographic image is pre-processed to enhance the quality of the image. In the second stage the welding region is extracted by using c-means segmentation method. After segmentation different features of the welded region are calculated and the 3-D contour plots are plotted..

Dental biometrics is used to identify the individuals in the forensic domain. Forensic identification is used to identify the person by comparing postmortem and ante mortem dental images. The dental radiograph image provides the information about teeth, tooth contours and in addition to that dental work (DW) is one of the notable issues for forensic identification. This dental work includes crown filling, root canal and crown mineralization for tooth. In this paper addresses the dental work identification for different radiographic images such as bitewing, periapical and panoramic dental images with Mahalanobis distance based matching. The proposed system contains preprocessing, feature extraction and matching. Experimental evaluation gives the satisfied results which will abet for dental biometrics.

Victim identification using dental radiographs is receiving much attention nowadays, since tooth is the more robust key component for forensic odontology. Missing tooth detection is one of the notable issues in dental identification system. Hence, this work proposes a novel method for identifying the persons by considering the missing tooth in addition with the contours of other teeth. Online region-based active contour model (ORACM) is used for shape extraction, and distance-based matching algorithm is used to identify the person by comparing both ante-mortem and post-mortem dental images. It concentrates on radiographs and photographs of dental images since adequate radiograph images may not be available for all circumstances. The proposed method is well suited for person identification, and it could assist forensic odontologists to identify the victims in an accurate manner.

Victim identification plays a vital role for identifying a person in major disasters at the time of critical situation when all the other biometric information was lost. At that time there is a less chance for identifying a person. The major issues of dental radiographs are dental work and missing or broken tooth was addressed in this paper. This algorithm can be established by comparing both ante mortem (AM) and post-mortem (PM) dental images. This research work is mainly focuses on the detection of dental work and broken tooth or missing tooth, then comparison of active contour model with mathematical model-based shape extraction for dental radiographic images are proposed. In this work, a new mathematical tooth approximation is presented and it is compared with online region-based active contour model (ORACM) is used for shape extraction. Similarity and distance-based technique gives better matching about both the AM and PM dental radiographs. Exact prediction of each method has been calculated and it is validated with suitable performance measures. The accuracy achieved for contour method is 94%, graph partition method is 96% and finally the hit rate of this method is plotted with cumulative matching characteristic (CMC) curve.

This paper highlights the effectiveness of Wavelet Octave COefficients of Residues (WOCOR) based feature extraction for robust text-independent speaker identification. A new feature set, WOCOR is proposed to capture the spectro temporal source excitation characteristics of the speech signal. This work is focused to increase the identification accuracy with databases containing short length speech signal. Experimental evaluation is carried out on TIMIT database with 630 speakers using Gaussian Mixture Model (GMM) is used as classifier. Vocal source feature is used to extract the information from the residual signal. The vocal source information contains pitch, pitch frequency and phase in the residual signal. In this project, 93.02% Identification rate is achieved in WOCOR.